Ashrey Ignise of Artus on Building What’s Next

What we talked about.

Sunny Ray talks with Ashrey Ignise of Artus about building a company around change intelligence, software that automates the impact analysis required when large organizations modify their products. Ignise explains that in industries like automotive and aerospace, assessing what a single change might break can take six to 12 months. He traces his path from the Uber Eats New Delhi launch team through founders office roles and consulting, where he often bridged business and engineering. Artus first launched as an AI product manager, but a feature on traceability and change intelligence drew unusually fast interest from a large automotive company, prompting a focused pivot. Ignise discusses why AI coding agents struggle across many repositories, how hidden dependencies slip through manual processes, and how a pilot is scoped around a single subsystem such as braking. He also describes the next 90 days, which center on tailoring the product to automotive and aerospace standards, and argues that faster AI-driven development makes change intelligence more important, since humans remain accountable for approving every change.

Impact analysis can take a year in automotive and aerospace. Ashrey Ignise explains how Artus makes it instant.

The questions, and the answers.

What are you building, and what problem does it solve?

We're building change intelligence. When a large company makes even a small change to a product, it takes a long time to figure out what will break. In automotive and aerospace, that impact analysis can take six to 12 months. Our software automates it, so you propose the change and instantly see its impact.

What first pulled you toward the problem of how one change ripples through an engineering organization?

In every company I worked at, I stood between business and engineering. Our first product was an AI product manager that turned business requirements into technical specifications. One feature, traceability and change intelligence, drew a huge response from a large automotive company, and deals moved faster than ever. That told me where to focus.

What did you do before Artus, and what do you still lean on today?

My first job was on the Uber Eats New Delhi launch team, onboarding restaurants in partnerships, which was sales from day one. Then I did founders office roles and consulting for tech startups. Across all of it, I was the bridge between what business wanted and what engineering executed, and that inspired our first product.

Why can't AI coding agents see the full impact of a change, and will better models fix that?

Coding agents are roughly 73% accurate on one repository, but research suggests that drops to about 35% around four or five repositories. Large automotive and aerospace companies have hundreds, plus hardware. Code can also behave differently at runtime than it reads, so tracking real impact takes deep, specialized work.

What was the moment the pivot to change intelligence clicked?

It was that automotive company. Our consumer product did well, with about 100 new signups a day in the first week, but enterprise sales lacked urgency. With change intelligence, champions quickly brought in decision makers and set next meetings. They also told us which tools had failed them, so we knew they'd pay today.

Which team feels this pain first, and how does a pilot work?

Usually the solutions architect, sometimes a product manager, who must produce a full impact analysis before leadership approves a change. A pilot targets one subsystem, like a braking system, and a change the company already made. They know the real blast radius, so they can compare manual time against Artus coverage and speed.

What do the next 90 days hold for Artus?

We're running pilots with a few big automotive companies. Change intelligence is not generic, because automotive and aerospace have their own processes, frameworks, and regulations. The next 90 days are about customizing our algorithm and workflow to embed deeply in those industry standards, which will make winning other companies in these spaces much easier.

As AI agents write more code, does change intelligence matter more or less?

It matters more. AI speeds up execution, so a company handling five changes a month might soon handle 20, 30, or 50. A human still has to be accountable and approve each one. Since agents work like a black box, tracking their changes and managing risk becomes far more important.